DSXchange

There are no easy general rules beyond the obvious. What kind of jobs are you talking about - database reads or writes or do you have a lot of transformations or special processing?

That's good enough for me - I was just checking to see if ODBC connectivity to the spreadsheet was possible. Have you checked the web for ODBC state "HY092" or native error 86 / 106?

Yes, you can use multiple filters. You might encounter runtime errors (I can't recall what they were) when using multiple pipes, but that can be solved by making the command "sh" and the arguments "'{your commands}'"

The latter part of your job, "pipe -> {processing} -> seq output" can be put into a separate DS job and both can be executed at the same time.

Are you successful in viewing the spreadsheet using the free Microsoft msquery or other tool that uses ODBC connectivity?

Server only has DATE and TIME and no TIMESTAMP.You would need to convert the date and time parts separately, then perform the computations, and finally convert and concatenate the result. In Stage variables: svInternalDate = ICONV(FIELD(In.TimeStampString,' ',1),'D4/MDY') svInternalT...

Gopalraja wrote:...Is it possible to unlock the Job without killing the Process ID in Unix or without restart the DS Server.

You can edit your $DSHOME/dsdlockd.config file and set "start=1", then start the deadlock daemon with the command

You can use an external filter and implement your choice of awk/sed or other program call that will only copy the part of the file you want or declare the file with just one big string record, run it through a transform to use substring commands to filter out the unwanted records, then use a column ...

Most likely your the width of the MQ Queue you are writing to is too small.

If you have multiple columns in real data records and only one column (no commas) in unwanted records you can use the reject row facility in the flat file stage to discard those.

In Server would work but I don't know if that applies to PX transforms. If not, use

The ".fs" is a user chosen ending. But the main drawback of a lookup fileset is that you cannot view the data. I tend to use a dataset for lookups, and then change over to a lookup fileset once I know everything is working correctly.

You can use "@PATH" in a BASIC routine.

manuel.gomez - just an additional tidbit of information regarding the PEEK stage. You can put it inline into a data stream, but you need to ensure that you make it use ALL data rows and just drag the input columns to output.

If you replace the switch stage and everything after that with a simple peek, do you still get the same error?